Earlier this week, I received an email from Stu:

Good morning, Alan!

"I just read one of the articles on your page about people souring on AI music, and it has me thinking. In the same vein as one of your OHONM episodes regarding the explosion of the music industry in the 80s and 90s before the inevitable downfall following the birth of Napster (and other P2P file-sharing services) and streaming, I'm curious on your thoughts towards the correlation of the growth in AI since then.

"My thoughts are that the small beneficial uses of AI (responsibly, that is) are quite limited to the services no longer affordable to most artists based on the current industry structure: without A&R reps hiring tons of talent and investing in their growth, new artists are left to their own devices and can't afford to pay for mix engineers or mastering technicians, so utilizing AI here seems initially appealing. The flip side, of course, is that if we just throw that portion of the industry to AI, nobody gets to hone their skills and then we lose the human touch there as well, which I admit is a major concern of mine in both my career outside of music as well as my children's futures.

"I guess my main question is this: Do you think that the greed of the music industry inevitably lead to its demise because it has forced people to other forms of access which in turn lead to the reduction in revenue for the companies so they stopped investing in the future artists which means many artists have to result to their boring day jobs just to survive and therefore less interest in pursuing a future in the industry... sort of a "fall of an empire" kind of scenerio? And whether the lack of "the dream" for coming generations leads to less people devoting their lives to developing the "skills that won't support a living wage", so just use AI to get your thoughts into songs because it's easier and there's no point wasting time on music anymore.

"I dunno, I guess I'm just feeling like music is dying a slow and painful death as I age into a cynical middle-aged ex-musician and wanted to vent to someone who is more committed and knowledgeable and has a more educated opinion towards not just AI ruining music, but the corporate greed ruining the industry for themselves as well...

Thanks for you time,
Stu

Here's my response.

Stu!

Excellent insight! We’re in the wild, wild west phase of AI music, which is analogous to how weird things got when sampling arrived in the middle 80s: No rules, no laws, no enforcement, lots of confusion as many people got ripped off. At the same time, sampling opened up brand new sonic vistas, sounds and approaches that were not only wilder than we imagined, but wilder than we COULD imagine. AI has the power to do the same thing, ethically, legally, and morally, with everything being fairly compensated.

The big difference today is that AI music is exponentially more powerful and thus harder to track and regulate. No one knows where this is going, although some people (including record labels) are laying big bets on their predictions. Their concern is that they’re not caught flat-footed like they were with file-sharing, digital downloads, and streaming. It’s possible that they’ll make mistakes that will be harmful to their businesses, and artists and fans as a result.

The old A&R guy trolling smoky clubs at 1 in the morning on a Tuesday is largely gone, replaced by data analytics sorting through streams, social media engagement, sales, and so on. But AI can’t FEEL anything. It doesn’t have a gut, that I-know-it-when-I-hear-it sensation. This will be a problem going forward.

AI platforms make music by guessing. Here’s a super-simplification.

Type in a prompt, and the program picks a starting note from trillions of data points scraped from songs created by real-life humans by its Large Language Model. Once it starts with that first note, it guesses what the third note should be. And then the fourth, and so on. In seconds, you end up with a song that statistically matches your prompt, stealing everything from musicians, many of whom can earn a living wage.

And the resulting AI music is PERFECT. No mistakes, beats quantized to the nanosecond, vocals perfectly tuned, no emotion, no accidents that lead to creative breakthroughs (think accidental guitar feedback at the beginning of The Beatles’ “I Feel Fine”), and no gear that goes on the fritz in the middle of a studio recording session.

Yes, creating with AI is cheaper and more efficient, but at what cost? We'll lose studios, producers, and engineers, but we'll lose more than that. Music's very soul will disappear.

Human failings and faults often make the best music. AI music makes people with no talent seem talented, and people with no creativity seem creative. It cheats those who spend 10,000 hours learning to get good. It just digs for gold while not creating anything original. AI will never give us the next Bob Dylan or Paul McCartney. We might get a copy of a copy of a copy, but it won't be real.

AI music is a solution to a problem that didn’t exist. I don’t want AI to create art. I want it to do my laundry and pick up after my dog, so I have more time to make art.

But there is room for optimism. We're already seeing Gen Z and Gen Alpha tire of living a virtual and digital life. It's possible that a backlash against AI will drive humans back into the arms of those who do music best: other humans.